JP5899075B2 - Automatic analyzer - Google Patents

Description

  The present invention relates to an automatic analyzer equipped with a sample dispensing device.

  In an automatic analyzer, for example, in a biochemical automatic analyzer, in order to analyze a component of a biological sample such as serum or urine, a test sample is reacted with a reagent, and a change in color tone or turbidity caused by the reaction is analyzed. Optically measured by a photometric unit such as a photometer.

  In order to cause the test sample and the reagent to react with each other, it is necessary to perform dispensing from the container in which the sample is accommodated to the reaction container. For this reason, the automatic analyzer is provided with a dispensing device that automatically sucks and discharges the sample or reagent from the container in which the sample or reagent is stored into the reaction container.

  In a sample dispensing apparatus that dispenses a test sample from a sample container to a reaction container, dispensing abnormality may occur due to various factors. The sample dispensing apparatus is provided with a probe for sucking and discharging the sample. However, a probe clogging due to suction of solid foreign matters such as fibrin is a frequent cause of dispensing abnormality. When the probe is clogged, a predetermined amount of sample cannot be dispensed into the reaction container, and a reliable analysis result cannot be obtained.

  In addition, when bubbles or liquid film exists on the liquid surface of the test sample, the bubble or liquid film is judged to be the liquid surface, the amount of sample to be sucked cannot be sucked, and dispensing abnormalities occur. To do.

  In order to avoid dispensing abnormalities when bubbles or liquid films exist, it is conceivable to increase the amount of probe immersed in the test sample. However, if the amount of probe immersed in the sample increases, contamination Nation increases, and the analysis result may be adversely affected.

  Therefore, in order to reduce the immersion depth of the probe in the liquid as much as possible, the liquid level of the liquid in the container is detected, and the descending operation of the probe is stopped when the tip of the probe reaches slightly below the liquid level. Then, it is common to control the operation so that a predetermined amount of liquid is sucked into the probe.

  As a means for detecting the liquid level of the test sample, the most commonly used capacitance change method is to detect a change in capacitance when the probe touches the liquid level. However, when such a liquid level sensor is used, as described above, if a bubble or a liquid film exists on the liquid level, it may be erroneously detected as the liquid level, resulting in dispensing abnormality.

  As a means to determine dispensing abnormalities as described above, many technologies have been proposed to detect dispensing abnormalities such as clogged sample probes based on pressure fluctuations by providing a pressure sensor in the dispensing flow path including the sample probe. ing.

  In Patent Document 1, data created from time-series data of pressure when suctioned normally is used as reference data, and time-series data of pressure sensor output values during suction is used as comparison data. A technique for detecting a suction abnormality based on the Mahalanobis distance calculated from the above is disclosed.

  According to the technique described in Patent Document 1, reference data is prepared in accordance with a dispensing condition including a dispensing amount, and an abnormality can be correctly detected even under different dispensing conditions. In addition, since abnormalities are detected by calculating the Mahalanobis distance based on time-series data of pressure over the entire interval from the start to the end of suction, not only abnormalities caused by specific causes but also abnormalities caused by various causes are detected. can do.

JP 2008-224691 A

  However, the method described in Patent Document 1 has a problem that it may be difficult to detect a dispensing abnormality under a dispensing condition with a small suction amount. For example, this is the case when minute foreign matter is sucked or when bubbles are sucked. In general, the pressure fluctuations in these cases tend to have a relatively small difference from the case of normal dispensing compared to the case where the probe is completely clogged or the case where a highly viscous sample is sucked.

  That is, under the condition of a small amount of suction, the suction / discharge operation time is short, and the pressure information that can be acquired therefrom is limited. For this reason, it is necessary to capture small changes from limited information, and it has been difficult to detect dispensing abnormalities.

  An object of the present invention is to realize an automatic analyzer capable of reliably detecting an abnormality during dispensing under a dispensing condition with a small amount of suction and performing a highly reliable analysis. In particular, in the second and subsequent dispensing operations of the same sample, the amount of suction is reduced as much as the dummy dose is not sucked, which may make it difficult to detect dispensing abnormality with high accuracy.

  In order to achieve the above object, the present invention is configured as follows.

  A typical invention is as follows.

  (1) The sample probe accommodated in the sample container is aspirated and discharged into the reaction container; the reagent accommodated in the reagent container is aspirated and discharged into the reaction container; and the reaction container accommodates The analysis unit for analyzing the sample, the pressure sensor for detecting the internal pressure of the sample probe, and the sample probe for the first time suction from the sample container and the suction pressure data of the sample discharged to the reaction container from the pressure sensor When the sample is aspirated from the sample container by the sample probe and discharged into the reaction container at the nth time (2 ≦ n (n) is a natural number) for the same sample, a plurality of normal values are obtained according to the suction pressure data. Normal abnormality determination reference data is selected from the abnormality determination reference data, and the nth same sample suction pressure data detected by the pressure sensor is detected. Calculating a characteristic variable, comparing the calculated characteristic variable with the normal / abnormal reference data, and determining whether or not there is an abnormality in the suction of the sample by the sample probe, the sample probe, the reagent probe, An automatic analyzer including the analysis unit and a controller that controls operations of the signal processing unit.

  (2) In another representative invention, the sample contained in the sample container is aspirated, the sample probe discharged to the reaction container, and the reagent contained in the reagent container are aspirated and discharged to the reaction container. The sample probe is first sucked from the sample container by the reagent probe, the analysis unit for analyzing the sample contained in the reaction container, the pressure sensor for detecting the internal pressure of the sample probe, and the sample probe. By discharging, the characteristic variable of the suction pressure data detected by the pressure sensor is calculated, and the same sample is aspirated from the sample container by the sample probe at the nth time (2 ≦ n (n) is a natural number). When discharging into the reaction container, the feature variable of the suction pressure data detected by the pressure sensor at the nth time is calculated, and the first feature variable and the nth feature variable are combined. Data as the determination target data for the nth time, normal / abnormality determination reference data according to the suction amount of the sample sucked for the first time, and normal / abnormality determination reference data according to the suction amount of the sample sucked at the nth time A signal processing unit that compares the reference data combined with the determination target data and determines whether or not there is an abnormality in the suction of the sample by the sample probe, the sample probe, the reagent probe, and the analysis unit And a controller for controlling the operation of the signal processing unit.

  (3) In another representative invention, the sample accommodated in the sample container is aspirated, the sample probe ejected to the reaction container, and the reagent accommodated in the reagent container are aspirated and ejected to the reaction container. The sample probe is first sucked from the sample container by the reagent probe, the analysis unit for analyzing the sample contained in the reaction container, the pressure sensor for detecting the internal pressure of the sample probe, and the sample probe. By discharging, the characteristic variable of the suction pressure data detected by the pressure sensor is calculated, and the same sample is aspirated from the sample container by the sample probe at the nth time (2 ≦ n (n) is a natural number). When discharging into the reaction container, the feature variable of the suction pressure data detected by the pressure sensor at the nth time is calculated, and the first feature variable and the nth feature variable are combined. Data as the determination target data for the nth time, the first normal / abnormal judgment reference data corresponding to the amount of the sample sucked for the first time, the suction amount of the nth sample and the pressure sensor In accordance with the detected suction pressure data, the reference data combining the second normal abnormality determination reference data selected from the plurality of normal abnormality determination reference data and the determination target data are compared, An automatic analysis comprising: a signal processing unit that determines whether or not there is an abnormality in the suction of the sample by the sample probe; and a controller that controls the operation of the sample probe, the reagent probe, the analysis unit, and the signal processing unit Device.

  To realize an automatic analyzer capable of reliably detecting an abnormality at the time of dispensing under a dispensing condition with a small amount of suction and performing a highly reliable analysis. In particular, an automatic analyzer capable of reliably detecting an abnormality and performing a highly reliable analysis in the second and subsequent dispensing operations of the same sample is realized.

1 is a schematic configuration diagram of an automatic analyzer to which the present invention is applied. It is explanatory drawing of the principal part (pressure signal processing part) in Example 1 of this invention. It is a figure which shows the internal state change of the sample probe from just before the start of aspiration to immediately after completion | finish. It is a rough form of the output of the pressure sensor at the time of sample suction. It is an internal block diagram of the signal processor in Example 1 of this invention. It is a figure which shows the example of calculation of the judgment amount in the judgment amount calculation part in Example 1 of this invention. It is a figure which shows the example of selection of the reference data by the reference data selection part in Example 1 of this invention. It is a figure which shows the example of the reference data used in order to calculate the Mahalanobis distance in Example 1 of this invention. It is a flowchart of the discrimination | determination operation | movement in Example 1 of this invention. It is a figure which shows the example of the amount of each time etc. in the actual analysis in Example 1 of this invention. It is a graph which compares the precision of abnormality determination when the present invention is applied and when the present invention is not applied. It is an internal block diagram of the signal processor in Example 2 of this invention. It is a flowchart of the discrimination | determination operation | movement in Example 2 of this invention. It is a figure which shows the example of the reference | standard data at the time of the nth dispensing in Example 2 of this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Example 1)
FIG. 1 is a schematic configuration diagram of an automatic analyzer to which the present invention is applied.

  Referring to FIG. 1, the automatic analyzer includes a sample disk (sample disk) 101 on which a plurality of sample containers (sample containers) 100 for holding a sample can be mounted, and a first reagent disk on which a plurality of reagent containers 103 for holding reagents can be mounted. 104 and a second reagent disk 106, and a reaction disk 111 in which a plurality of reaction vessels 110 are arranged on the circumference.

  The automatic analyzer also supplies a sample probe (sample probe) 102 for dispensing a sample sucked from the sample container 100 to the reaction container 110 and a reagent sucked from the reagent container 103 in the first reagent disk 104 to the reaction container 110. A first reagent probe 105 to be dispensed and a second reagent probe 107 to dispense the reagent aspirated from the reagent container 103 in the second reagent disk 106 into the reaction container 110 are provided.

  Further, the automatic analyzer includes a stirring device 108 for stirring the liquid in the reaction vessel 110, a container cleaning mechanism 109 for washing the reaction vessel 110, a light source 112 installed near the inner periphery of the reaction disk 111, and spectral detection. 113, a computer 114 connected to the spectroscopic detector 113, and a controller 115 that controls the operation of the entire automatic analyzer and exchanges data with the outside.

  The sample probe 102 is connected to the metering pump 117 via the dispensing channel 116, and a pressure sensor 118 is provided in the middle of the dispensing channel 116.

  FIG. 2 is an explanatory diagram of a main part (a pressure signal processing unit in the sample probe 102) according to the first embodiment of the present invention.

  In FIG. 2, a narrowed portion 200 having a smaller cross-sectional area than other portions is formed at the tip of the sample probe 102. The metering pump 117 is provided with a plunger 202 driven by a drive mechanism 201. The metering pump 117 is connected to the pump 204 through the valve 203.

  The pressure sensor 118 is connected to a signal processor (signal processing unit) 206 via an AD converter 205. The sample probe 102 is filled with the system liquid 207 and the sample 209 is sucked through the separation air 208. The example shown in FIG. 2 shows a state where the sample 209 is sucked into the sample probe 102.

  The sample probe 102 has a moving mechanism (not shown), and can move up and down and rotate to the sample container 100 and the reaction container 110.

  Next, the operation in the first embodiment of the present invention will be described.

  1 and 2, a sample to be examined such as serum is placed in a sample container 100 and set on a sample disk 101. The type of analysis required for each sample is input from the computer 114 to the controller 115. A sample collected from the sample container 100 by the sample probe 102 is dispensed into the reaction container 110 arranged on the reaction disk 111.

  Then, a certain amount of reagent is dispensed from the reagent container 103 installed on the first reagent disk 104 or the second reagent disk 106 to the reaction container 110 by the first reagent probe 105 or the second reagent probe 107, and is supplied to the stirring device 108. And stirred. The sample and reagent dispensing amounts are set in advance for each type of analysis.

  The reaction disk 111 is periodically rotated and stopped, and photometry is performed by the spectroscopic detector 113 at the timing when the reaction vessel 110 passes in front of the light source 112. Photometry is repeated by the spectroscopic detector 113 during the reaction time of 10 minutes, and then the reaction liquid in the reaction vessel 110 is discharged and washed by the vessel washing mechanism 109. In the meantime, in another reaction vessel 110, operations using different samples and reagents are performed in parallel. Data measured by the spectroscopic detector 113 is calculated by the computer 114, and the concentration of the component corresponding to the type of analysis is calculated and displayed on the display of the computer 114.

  The operation of the sample probe 102 will be described in detail.

  Before the sample is sucked, first, the controller 115 opens and closes the valve 203 to fill the flow path of the sample probe 102 with the system liquid 207 supplied from the pump 204. Next, the controller 115 lowers the plunger 202 by the drive mechanism 201 while the tip of the sample probe 102 is in the air, and sucks the separation air 208.

  Next, the controller 115 lowers the sample probe 102 into the sample container 100, and lowers the plunger 202 by a predetermined amount with its tip immersed in the sample to suck the sample into the probe 102. In this case, the suction liquid 209 is a sample. The pressure fluctuation of the sample probe 102 during the operation following the suction is detected by the pressure sensor 118, converted into a digital signal by the AD converter 205, and sent to the signal processor 206. Thereafter, the sample probe 102 is moved onto the reaction vessel 110 to discharge the sample.

  The pressure fluctuation during the operation following the sample discharge of the sample probe 102 is detected again by the pressure sensor 118, converted into a digital signal by the AD converter 205, and sent to the signal processor 206. Subsequently, the inside and outside of the sample probe 102 is cleaned by opening and closing the valve 203 to prepare for the next analysis.

  The signal processor 206 determines the presence or absence of dispensing abnormality from the pressure waveforms at the time of sample aspiration and discharge by the sample probe 102. If it is determined that there is an abnormality, the analysis is stopped and the computer 114 displays An alarm is displayed on the part, etc., and the recovery operation is performed. The return operation is selected from among other things such as re-dispensing except for the cause of the abnormality, moving to inspection of another sample, and stopping the apparatus.

  FIG. 3 is a diagram showing a change in the internal state of the sample probe 102 from immediately before the start of suction to immediately after the end of suction.

  When there are a plurality of analysis item measurement requests for the same sample, the sample is dispensed according to the number of analysis items. In FIG. 3, in the initial dispensing of the same sample, in addition to the discharge dose 209b, a dummy dose 209a is aspirated for the purpose of preventing thinning of the sample. In the second and subsequent dispensing, the suction and discharge operations are repeated with the dummy dose 209a held in the sample probe 102. Usually, 6 to 12 μl of suction is performed as a dummy dose. Since the first dose is aspirated with the dummy dose 209a, it is easier to catch an abnormality during the dispense from the pressure fluctuation during the aspiration compared to the second and subsequent doses even with the same dispense amount. On the other hand, in the second and subsequent dispensing, since there is no suction of the dummy dose 209a, it is difficult to catch an abnormality at the time of dispensing from the pressure fluctuation at the time of suction. In the present invention, the pressure fluctuation information obtained at the time of the first dispensing is reflected in the abnormality determination after the second time, and the abnormality detection performance after the second time is improved.

  FIG. 4 is a schematic view of the output of the pressure sensor 118 during sample suction.

  In FIG. 4, the vertical axis represents the pressure due to the output of the pressure sensor 118 and the horizontal axis represents the time axis. In the waveform shown in FIG. 4, the solid line indicates the pressure fluctuation when the low-viscosity sample A (within the normal range) and the broken line indicates the high-viscosity sample B (within the normal range). It can be seen that the amount of change in pressure and the appearance of vibrations differ depending

  As shown in FIG. 4, the higher the viscosity of the sample B, the larger the amount of change from the pressure value before the suction start to the pressure value during the suction, and the vibration immediately after the start of suction and immediately after the end is suppressed. Even when a normal sample is sucked in this way, there are various pressure fluctuation patterns. In particular, when aspiration is performed with a small amount of suction, the amount of information that can be acquired is reduced because the suction section is shortened. Therefore, it may be difficult to distinguish between the suction of bubbles in the high-viscosity sample B in the normal range and the suction of the low-viscosity sample A in the normal range when fine foreign matter is sucked. However, when performing highly accurate abnormality detection, it is desirable to determine which one or both are abnormal. For this reason, the viscosity of the sample is estimated to some extent by the first suction, and in the second and subsequent sample dispensing, optimum normal / abnormality determination reference data is selected according to this viscosity, and abnormality determination is performed. Details will be described later.

  The reason for estimating the viscosity in the first suction is that there is a dummy dose suction, so that it is larger than the actual discharge dose, and the suction section is longer than when there is no dummy dose suction. Therefore, this is because it is suitable for estimating the differential pressure described later, that is, the viscosity of the sample to be sucked.

  FIG. 5 is an internal configuration diagram of the signal processor 206 according to the first embodiment.

  In FIG. 5, the signal processor 206 includes a determination amount calculation unit 206e, a reference data selection unit 206f, a statistical distance calculation unit 206a, a memory 206b, a statistical distance calculation unit 206a, a comparison unit 206c, and a determination unit 206d. Is provided. The determination unit 206 d of the signal processor 206 transmits the determination result to the controller 115. The signal processor 206 may be provided separately from the controller 115, or may be provided in the controller 115.

  FIG. 6 is a diagram illustrating a calculation example of the determination amount in the determination amount calculation unit 206e. In FIG. 6, P0 is the output of the pressure sensor 118 when the sample has not been sucked yet immediately after the liquid level is detected. The output of the pressure sensor 118 in this state is a value affected by the atmospheric pressure.

  In FIG. 6, P1 is the output of the pressure sensor 118 immediately before the end of suction. By obtaining the differential pressure ΔP between P0 and P1, the influence of atmospheric pressure can be avoided. In a dispensing amount range in which the drive pattern of the drive mechanism 201 that is a motor is the same, the differential pressure ΔP between P0 and P1 reflects the viscosity of the sample. Since the output value of the pressure sensor 118 usually fluctuates finely, it is desirable that P0 and P1 are average values in a certain section.

  FIG. 7 is a diagram illustrating an example of selection of reference data in the reference data selection unit 206f. In FIG. 7, the reference data selection unit 206f uses the determination amount calculated by the determination amount calculation unit 206a to select from a plurality of reference data held.

  In this example, three reference data are prepared, and the reference data is selected based on the judgment amount. Each reference data is prepared for each dispensing amount or for each fixed dispensing amount range as a result of the characteristic variable.

  When the differential pressure is Pa or more and less than Pb, the reference data I is obtained, and the target viscosity is 0.9 to 1.6 mPa · s. Moreover, when the differential pressure is Pb or more and less than Pc, the reference data II is obtained, and the target viscosity is 1.6 to 2.3 mPa · s. Further, when the differential pressure is equal to or higher than Pc and lower than Pd, the reference data III is obtained, and the target viscosity is 2.3 to 3.0 mPa · s.

  The statistical distance D calculated by the statistical distance calculation unit 206a is an index obtained by quantifying the similarity between two events represented by a plurality of feature variables. In the case of the first embodiment of the present invention, the reference data This is a calculation of how far the data measured at that time is. There are various calculation methods for the statistical distance D, such as Mahalanobis distance, Euclidean distance, standard Euclidean distance, Manhattan distance, Chebyshev distance, Minkowski distance, and multivariate normal density.

  For example, Mahalanobis distance D is calculated as follows.

  FIG. 8 is a diagram illustrating an example of reference data used to calculate the Mahalanobis distance in the first embodiment of the present invention.

As shown in FIG. 8, if the characteristic variable number is 1 to k and the event number is 1 to n, the reference data is x _{11 to} x _1k ,... against n, reference _data, and x n1 _{~x nk.}

Each feature variable _{x 1,} x 2 to be determined, ..., to _{x k,} the average x of each feature variable reference data _'1, x' _2, ..., x _'k (Equation (1) the variable subjecting the bar to x) and standard deviation σ _1, σ _{2, ···,} using sigma _k, normalized performs calculation of the following equation (1).

  However, in Formula (1), i = 1, 2, ..., k (integer).

  A correlation matrix between each feature variable is obtained from the reference data, and the following calculation is performed from the inverse matrix A and the feature variable group normalized by the following equation (2) to calculate the Mahalanobis distance D.

  FIG. 9 is a flowchart of the determining operation in the first embodiment of the present invention.

  In FIG. 9, the sample probe 102 performs a suction operation in the first dispensing (STEP 1), and the statistical distance calculation unit 206a calculates a characteristic variable for the digital signal of the pressure waveform sent from the AD converter 205. (STEP2).

  The characteristic variable here may be anything that can accurately represent the pressure fluctuation pattern. For example, the pressure average value at regular time intervals and the appearance timings of the minimum and maximum points of pressure fluctuation that appear when the plunger 202 starts and stops operation.

  Further, the signal processor 206 holds the result of the characteristic variable when the sample probe 102 has successfully dispensed the sample as reference data in the memory 206b, and the statistical distance calculation unit 206a stores the calculated characteristic variable A statistical distance D from the reference data is calculated (STEP 3).

  Then, the comparison unit 206c compares the statistical distance D with a predetermined threshold value stored in the memory 206b (a threshold value for determining whether or not there is an abnormality in the sample suction by the sample probe) (STEP 4). The result is supplied to the determination unit 206d.

  When the statistical distance D is greater than the predetermined threshold, the determination unit 206d determines that there is an abnormality in the suction, and performs the return process 1 (STEP 5). In the return process 1, the determination unit 206d transmits an abnormality to the controller 115 and the computer 114, and the operation proceeds to the alarm process such as an alarm display and the next sample process on the display screen (abnormality notification unit) of the computer 114. This process is executed under the control of the controller 115 and the computer 114.

  In STEP 4, when the statistical distance D is smaller than the threshold value, the first ejection operation is performed under the control of the controller 115 in accordance with a command from the determination unit 206d (STEP 6). Next, it is determined whether or not the second and subsequent dispensing has been ordered (STEP 7). If the second and subsequent dispensing has not been ordered, the process ends. In STEP 7, when the second and subsequent dispensing is ordered, the controller 115 determines whether or not the determination can be made independently based on the minimum amount (described later) that can determine whether or not the dispensing is correct (STEP 8). The minimum amount that can be determined for dispensing correctness is stored in advance in the memory 206b as a threshold for dispensing amount, and is determined based on whether the planned suction amount is greater than or less than (or exceeds or less than) this threshold. In STEP 8, if it is possible to make an independent determination (greater than or equal to the threshold value), the controller 115 instructs the signal processor (signal processing unit) 206 to select the reference data that meets the dispensing amount condition at that time. (STEP 10), then proceed to STEP 11. The reason for performing the determination of STEP 8 is that when there is a large amount of dispensing after the second time, it is not necessary to select one of a plurality of reference data in the same dispensing amount, for example, reference data I, II, III. This is because it is possible to determine whether or not there is a suction abnormality by using reference data that meets predetermined dispensing volume conditions.

  In STEP 8, if the determination by itself is not possible (less than the threshold), the reference data selection unit 206f follows the command from the controller 115, and the reference data selection unit 206f determines the dispensing amount and the determination amount (differential pressure ΔP used in the first suction operation). ) To select reference data (STEP 9).

  Next, the second and subsequent suction operations of the same sample are performed (STEP 11). When it is determined in step STEP10 that the reference data is selected, the reference data that meets the dispensing amount condition is selected by the determination amount calculation unit 206e and the reference data selection unit 206f. Then, the statistical distance calculation unit 206a calculates (extracts) the characteristic variable for the digital signal of the pressure waveform sent from the AD converter 205 (STEP 12). The statistical distance calculation unit 206a calculates a statistical distance D from the selected reference data of the calculated feature variable (STEP 13). Then, the comparison unit 206c compares the statistical distance D with a predetermined threshold stored in the memory 206b, and supplies the result to the determination unit 206d to determine whether the statistical distance D is smaller than the threshold. (STEP 14).

  If the statistical distance D is greater than or equal to the predetermined threshold, the determination unit 206d determines that there is an abnormality in the suction, and performs the return process 2 (STEP 15). The return process 2 is a process in which the determination unit 206d transmits the presence of an abnormal suction to the controller 115 and the computer 114, and executes an operation of proceeding to the alarm process and the next sample process under the control of the controller 115 and the computer 118.

  In step STEP14, when the statistical distance D is smaller than the threshold value, the operation is shifted to the discharge operation under the control of the controller 115 in accordance with a command from the determination unit 206d (STEP16). Thereafter, STEP 7 to STEP 16 are repeated until the next dispensing is completed.

  FIG. 10 is an example in which the first embodiment of the present invention is applied in actual analysis, and uses the initial information on the dispensing amount, the dummy amount, the discharge capacity, whether or not the determination can be made independently. It is a figure which shows the example of whether to determine. In FIG. 10, an amount (the minimum amount that can be determined for dispensing correct / incorrect) that can be independently determined to determine whether the dispensing is correct or not (STEP 8 in FIG. 9) is experimentally determined, and this dispensing amount is set in advance. .

  In this case, it is assumed that 4.0 μl is set as the minimum amount that can be used to determine whether or not dispensing is correct. In this case, the dummy dose is 10.0 μl. In the first dispensing, a dummy dose is aspirated in addition to the discharge dose of 2.0 μl. When it is determined that it is normal or not, the differential pressure ΔP is stored in the memory 206b for selection of reference data used for the determination of correctness of the second and subsequent dispensing. deep.

  In the second dispensing, suction is performed with a discharge volume of 1.2 μl. Since it is less than the minimum amount of 4.0 μl that can be used to determine whether dispensing alone or not, the determination is made using the initial information. The determination here is performed using reference data selected from the differential pressure ΔP at the first dispensing and the second dispensing amount.

  In the third dispensing, a suction volume of 8.0 μl is aspirated. Since the minimum amount that can be determined for dispensing correctness alone is 4.0 μl or more, the determination using the first information is not necessary, and the correctness determination here is performed independently. In the fourth dispensing, suction of 3.5 μl of discharge volume is performed. Since this is less than the minimum amount of 4.0 μl at which dispensing correctness can be determined, the determination is made using the initial information. The determination here is performed using reference data selected from the differential pressure ΔP at the time of the first dispensing and the fourth dispensing amount.

  FIG. 10 shows an example in which the minimum amount that can determine whether dispensing is correct or not is set. However, it is not always necessary to set such an amount, and the reference data is uniformly selected based on the judgment amount (STEP 9 in FIG. 9). ) That is, STEP 8 in FIG. 9 may be omitted. However, setting this minimum amount eliminates the need to store a plurality of reference data for each dispensing amount in a dispensing amount larger than this minimum amount in the memory 206b. For example, the memory 206b The capacity of can be reduced.

  FIG. 11 is a graph comparing the accuracy of abnormality determination when the present invention is applied and when the present invention is not applied. (A) of FIG. 11 is an example in the case where the present invention is not applied, and shows an example in which the information of the first dispensing is not reflected, and (b) in FIG. 11 is an example in the case of applying the present invention. An example of reflecting the information of the first dispensing is shown.

  In FIG. 11 (a), there is an overlapping portion between the normal maximum distance and the minimum minimum distance in the statistical distance, and it is determined to be normal while being abnormally dispensed. There are cases where it is judged that there is.

  On the other hand, when the present application shown in FIG. 11B is applied, there is no overlap between the normal maximum distance and the abnormal minimum distance in the statistical distance, and the abnormal dispensing and the normal dispensing are clearly defined. It can be seen that there is a distinction.

  As described above, according to the first embodiment of the present invention, even in the second and subsequent dispensing conditions with a small amount of suction, an automatic operation capable of reliably detecting an abnormality during the dispensing and performing a highly reliable analysis. An analyzer can be realized.

In the first embodiment, the application to the pressure fluctuation at the time of suction has been described. However, the present invention can be similarly applied to the pressure fluctuation at the time of discharge.
(Example 2)
Next, a second embodiment of the present invention will be described. In addition, about a structure and operation | movement common to Example 1, illustration and the detailed description are abbreviate | omitted. That is, the configurations and operations shown in FIGS. 1 to 3 described in the first embodiment are the same configurations and operations in the second embodiment.

  The difference between the second embodiment and the first embodiment of the present invention is that, in the first embodiment, reference data suitable for determining the correctness of the second and subsequent dispensing is selected based on the information from the pressure fluctuation during the first suction operation. On the other hand, in the second embodiment, whether or not the correctness determination is performed for the second and subsequent times by calculating the statistical distance between the reference data and the determination target, The correctness of dispensing is determined from information obtained by combining the characteristic variables of the pressure fluctuation during the subsequent suction operation.

  FIG. 12 is an internal configuration diagram of the signal processor 206 according to the second embodiment of the present invention. In FIG. 12, the signal processor 206 includes a statistical distance calculation unit 206a, a memory 206b, a comparison unit 206c, a determination unit 206d, and a reference data selection unit 206f.

  The determination unit 206 d of the signal processor 206 transmits the determination result to the controller 115. The signal processor 206 may be provided separately from the controller 115 or may be provided in the controller 115. When the suction operation in the first dispensing is performed, the statistical distance calculation unit 206a calculates a characteristic variable for the digital signal of the pressure waveform sent from the AD converter 205. The characteristic variable here may be anything that can accurately represent the pressure fluctuation pattern. For example, the pressure average value at regular time intervals and the appearance timings of the minimum and maximum points of pressure fluctuation that appear when the plunger 202 starts and stops operation.

  Then, a statistical distance between the reference data stored in the memory 206b and the calculated feature variable is calculated. In the case of the first suction operation, the reference data selection unit 206f instructs the statistical distance calculation unit 206a to select prestored reference data from the memory 206 according to the suction amount.

  When the sample probe 102 has successfully dispensed the sample, the result of the characteristic variable at that time is stored as data in the memory 206b. In the second suction, the result of the characteristic variable is stored as data in the memory 206b. The determination target whether or not the second normal suction operation is performed is data obtained by combining the result of the first feature variable and the result of the second feature variable. As the reference data, the statistical distance of the combination data of the above characteristic variables is calculated by using, as the reference data, data obtained by combining the reference data determined according to the first dispensing amount and the reference data determined according to the second dispensing amount. Determine whether normal suction.

  In the n-th (2 ≦ n (n) is a natural number) sample aspiration, the combination of the first aspiration feature variable stored in the memory 206b and the n-th aspiration feature variable is a determination target. Normality / abnormality determination is performed by using, as reference data, data obtained by combining reference data determined in accordance with the initial dispensing amount and reference data determined in accordance with the nth dispensing amount.

  FIG. 13 is a flowchart of the determination operation in the second embodiment of the present invention. The difference between the determination operation flowchart in the second embodiment and the determination operation flowchart in the first embodiment is that step STEP9 in the first embodiment is replaced with step STEP9A in the second embodiment.

  That is, in step STEP9A of FIG. 13, data that is a combination of the characteristic variable at the time of normal first suction and the characteristic variable at the time of the nth suction is determined as the determination target, and is determined according to the first dispensing amount. The reference data consisting of a combination of the reference data and the reference data determined according to the nth dispensing amount is selected. The other steps are the same as in the first and second embodiments. In STEP 8, when the determination by itself is possible (STEP 10), the reference data and the data to be determined do not need to be combined data as described above, and are the same as in the first embodiment. Therefore, depending on the determination in STEP 8, the initial reference data in STEP 12 and 13 and the handling of the first characteristic variable are different.

  FIG. 14 is a diagram illustrating an example of combined reference data at the time of the n-th dispensing in the second embodiment of the present invention. It can be seen that by including the first reference data in the reference data, the amount of information is larger than that of the nth reference data alone. Therefore, in the second and subsequent dispensing conditions with a small amount of suction, the decrease in the amount of information based on the reduction in the amount of suction can be supplemented with the first reference data, and abnormalities during dispensing can be detected reliably. Realize an automated analyzer that can perform highly reliable analysis.

  As described above, also in the second embodiment of the present invention, the same effect as in the first embodiment can be obtained.

(Example 3)
Next, Embodiment 3 of the present invention will be described. The third embodiment is a combination of the first and second embodiments. As in the first embodiment, based on the pressure fluctuation information at the time of the first suction operation, the reference data suitable for the correctness determination for the nth dispensing (n ≧ 2) is selected from a plurality of reference data held in advance. . Furthermore, the reference data selected here is data composed of a combination of the reference data at the first suction operation and the reference data selected previously, as in the second embodiment. On the other hand, the determination target is a combination of the characteristic variable of the first suction pressure data and the characteristic variable of the nth suction pressure data, as in the second embodiment.

  In the correctness determination of the n-th time (n ≧ 2), if the determination can be made independently, the reference data that matches the dispensing amount condition is selected. When it cannot be used alone, the characteristic variable calculated from the pressure fluctuation at the first suction and the characteristic variable calculated from the pressure fluctuation at the n-th suction are combined to form one pattern data. Then, a statistical distance between the pattern data and the selected reference data is calculated. Whether the dispensing is correct or not is determined by comparing the calculated statistical distance with a threshold value.

  The internal configuration and other configurations of the signal processing unit 206 in the third embodiment are the same as those in the second embodiment.

  In the third embodiment of the present invention, the same effect as in the first and second embodiments can be obtained.

  In addition, this invention is applicable not only to a biochemical automatic analyzer but also to an immune automatic analyzer.

  Further, the present invention can be similarly applied not only to pressure fluctuation during suction but also to pressure fluctuation during discharge.

  DESCRIPTION OF SYMBOLS 100 ... Sample container, 101 ... Sample disk, 102 ... Sample probe, 103 ... Reagent container, 104 ... First reagent disk, 105 ... First reagent probe, 106 ... Second reagent disk 107 ... Second reagent probe 108 ... Stirrer 109 ... Container cleaning mechanism 110 ... Reaction vessel 111 ... Reaction disk 112 ... Light source 113 ... Spectral detector, 114 ... Computer, 115 ... Controller, 116 ... Dispensing flow path, 117 ... Quantitative pump, 118 ... Pressure sensor, 200 ... Throttle part, 201 ... Drive mechanism, 202 ... Plunger, 203 ... Valve, 204 ... Pump, 205 ... A / D converter, 206 ... Signal processor, 206a ... judgment amount calculation unit, 206b ... memory, 206c ... comparison unit, 206d ... determination unit, 206e ... judgment amount calculation unit, 206f ... reference data calculation unit 207: System liquid 208: Separated air 209: Suction liquid

Claims (12)

A sample probe for aspirating the sample contained in the sample container and discharging it to the reaction container;
A reagent probe that aspirates the reagent contained in the reagent container and discharges it to the reaction container;
An analysis unit for analyzing the sample stored in the reaction container;
A pressure sensor for detecting the internal pressure of the sample probe;
With the sample probe, the suction pressure data of the sample that is sucked from the sample container for the first time and discharged to the reaction container is acquired from the pressure sensor,
When the same sample is suctioned from the sample container by the sample probe and discharged to the reaction container at the nth time (2 ≦ n (n) is a natural number), a plurality of normal / abnormal judgment criteria are determined according to the suction pressure data. Select normal / abnormal judgment criteria data from the data,
Calculating a characteristic variable of the n-th same sample suction pressure data detected by the pressure sensor by the sample probe;
A signal processing unit that compares the calculated feature variable with the normal / abnormal reference data and determines whether or not there is an abnormality in the suction of the sample by the sample probe;
A controller that controls operations of the sample probe, the reagent probe, the analysis unit, and the signal processing unit;
An automatic analyzer characterized by comprising. The automatic analyzer according to claim 1,
The automatic analyzer according to claim 1, wherein the normal / abnormal determination criterion data is selected according to a pressure difference between a pressure in the sample probe of the suction pressure data and a pressure in the sample probe immediately before the end of sample suction. The automatic analyzer according to claim 2,
The signal processing unit sucks from the sample container by the sample probe at the nth time (2 ≦ n (n) is a natural number) and discharges to the reaction container at the nth time for the same sample, If the amount of the sample to be discharged is less than a predetermined amount, the normal / abnormal judgment criterion data is selected according to the pressure difference of the first sample information and the suction amount of the nth sample,
If the amount of the sample to be sucked and discharged is equal to or greater than the predetermined amount, normal / abnormal judgment reference data is selected according to the amount of the sample to be sucked and discharged,
The feature variable of the suction pressure data at the nth time of the same sample by the sample probe is compared with any of the selected reference data, and it is judged whether or not there is an abnormality in the suction of the sample by the sample probe, Automatic analyzer to do. The automatic analyzer according to claim 3,
The signal processor is
A determination amount calculation unit that calculates a determination amount that is a pressure difference between the suction pressure data of the sample by the sample probe;
By comparing the reference data with the characteristic variables and reference data of normal / abnormal judgment criteria data predetermined for each sample suction amount, initial sample suction pressure data by the sample probe, and sample suction pressure data by the sample probe A memory for storing an abnormality determination threshold for determining whether or not there is an abnormality in the suction of the sample by the probe;
A reference data selection unit that selects one of the plurality of normal / abnormal determination criterion data stored in the memory based on the determination amount calculated by the determination amount calculation unit;
A statistical distance calculator for calculating a statistical distance between the characteristic variable of the suction pressure data of the sample by the sample probe and any of the selected reference data;
A comparison unit that compares the statistical distance calculated by the statistical distance calculation unit with the abnormality determination threshold value stored in the memory;
A determination unit for determining whether the suction operation of the sample by the sample probe is normal or abnormal based on the comparison by the comparison unit;
The automatic analyzer characterized by having. The automatic analyzer according to claim 4,
The pressure difference is the detected pressure of the pressure sensor between the time when the sample probe detects the liquid level of the sample and stops and the time when the sample is aspirated, and the pressure detected by the pressure sensor during the aspiration of the sample. This is an automatic analyzer characterized by the differential pressure. The automatic analyzer according to claim 4,
The automatic analyzer according to claim 1, wherein the normal / abnormal judgment reference data stored in the memory is pressure waveform data classified according to the viscosity of the sample. The automatic analyzer according to claim 4,
The automatic analysis apparatus characterized in that the statistical distance calculated by the statistical distance calculation unit is any one of Mahalanobis distance, Euclidean distance, standard Euclidean distance, Manhattan distance, Chebyshev distance, Minkowski distance, and multivariate normal density. A sample probe for aspirating the sample contained in the sample container and discharging it to the reaction container;
A reagent probe that aspirates the reagent contained in the reagent container and discharges it to the reaction container;
An analysis unit for analyzing the sample stored in the reaction container;
A pressure sensor for detecting the internal pressure of the sample probe;
A characteristic variable of the suction pressure data detected by the pressure sensor is calculated by first suctioning from the sample container and discharging to the reaction container by the sample probe, and the nth (2 ≦ n (n ) Is a natural number), when the sample probe is sucked from the sample container and discharged to the reaction container, the characteristic variable of the suction pressure data detected by the pressure sensor is calculated for the nth time, and the first characteristic variable is calculated. And the nth feature variable are calculated as the determination target data for the nth time,
Reference data combining normal / abnormal judgment reference data according to the amount of suction of the sample aspirated for the first time and normal / abnormal judgment reference data according to the suction amount of the sample aspirated n-th time; A signal processing unit that determines whether or not there is an abnormality in the suction of the sample by the sample probe,
A controller that controls operations of the sample probe, the reagent probe, the analysis unit, and the signal processing unit;
An automatic analyzer characterized by comprising. The automatic analyzer according to claim 8,
When the signal processing unit sucks from the sample container by the sample probe and discharges to the reaction container at the nth time (2 ≦ n (n) is a natural number) in the same sample, the signal processing unit sucks at the nth time, If the amount of the sample to be discharged is less than a predetermined amount, the combined reference data is compared with the combined determination target data,
If the amount of the sample to be aspirated and discharged is equal to or greater than the predetermined amount, the normal / abnormal judgment reference data according to the aspirated amount of the sample to be aspirated and discharged and the suction pressure data at the nth time of the same sample by the sample probe Compare feature variables as judgment target data,
An automatic analyzer which judges whether or not there is an abnormality in the suction of the sample by the sample probe. The automatic analyzer according to claim 9,
The signal processor is
An abnormality determination threshold for determining whether or not there is an abnormality in the suction of the sample by the sample probe;
A statistical distance calculator for calculating a statistical distance between the determination target data and the normal / abnormal determination criterion data;
A comparison unit that compares the statistical distance calculated by the statistical distance calculation unit with the abnormality determination threshold value stored in the memory;
A determination unit for determining whether the suction operation of the sample by the sample probe is normal or abnormal based on the comparison by the comparison unit;
The automatic analyzer characterized by having. A sample probe for aspirating the sample contained in the sample container and discharging it to the reaction container;
A reagent probe that aspirates the reagent contained in the reagent container and discharges it to the reaction container;
An analysis unit for analyzing the sample stored in the reaction container;
A pressure sensor for detecting the internal pressure of the sample probe;
A characteristic variable of the suction pressure data detected by the pressure sensor is calculated by first suctioning from the sample container and discharging to the reaction container by the sample probe, and the nth (2 ≦ n (n ) Is a natural number), when the sample probe is sucked from the sample container and discharged to the reaction container, the characteristic variable of the suction pressure data detected by the pressure sensor is calculated for the nth time, and the first characteristic variable is calculated. And the nth feature variable are calculated as the determination target data for the nth time,
In accordance with the first normal / abnormal judgment reference data corresponding to the amount of sample sucked for the first time, the amount of sample sucked n times and the suction pressure data detected by the pressure sensor, a plurality of normal Whether or not there is an abnormality in the suction of the sample by the sample probe by comparing the determination target data with the reference data combining the second normal abnormality determination reference data selected from the abnormality determination reference data A signal processing unit for determining
A controller that controls operations of the sample probe, the reagent probe, the analysis unit, and the signal processing unit;
An automatic analyzer characterized by comprising. The automatic analyzer according to claim 11,
The second normal / abnormal judgment reference data is selected in accordance with a pressure difference between the pressure in the sample probe of the suction pressure data and the pressure in the sample probe immediately before the end of sample suction. Analysis equipment.

Images